A place to post ridiculous posts from linkedIn.com
(Full transparency.. a mod for this sub happens to work there.. but that doesn't influence his moderation or laughter at a lot of posts.)
all of those positions have junior roles. what is he talking about?
"junior doctors who learned to diagnose through youtube", another litmus paper here: a person with no expertise or experience in a field making grand claims about it. this guy seems like full on bullshitter, perhaps out of all the professions he counted, he is the one most easily replaceable by LLMs
Yup, doctors need to go through residency and get their first job, just like a software engineer usually needs to get an internship and then their first job.
This isn’t a lunatic. This is someone trying to make a point about companies thinking they can use AI to replace devs. Poe’s Law is on heavy display here in these comments.
Whether or not you have experienced it, there is currently a trend both in recruiting and in millionaire leadership dialogue toward dropping devs for AI codegen. CEOs that don’t understand how anything works (eg Salesforce) think you can just not hire devs because Google’s inflated AI stats that included basic autocomplete in their full AI codegen numbers indicate AI can code. Boards believe generative AI is capable of things it won’t be able to touch for decades. I have to deal with idiotic AI questions from Fortune 500 companies every fucking week.
From a hiring perspective, it’s becoming incredibly difficult to weed out AI bullshit. For every one qualified candidate I get, I’ve had to drop five or more in a fucking tech screen because, while codegen has given them enough to pass a basic hiring screen that used to weed out a lot more, there’s zero fucking ability to code without Copilot or critical understanding of the code it generates. When I was starting out, the same problem existed at university but got filtered out after graduation fairly quickly.
The non lunatic here is extending that to other disciplines because it’s a natural next question. He’s not exactly applying a slippery slope; it’s sort of there underneath.
Edit: valid criticism of the post is that you have to have a degree to code. That’s bullshit. After my first degree, I went back for CS and dropped out because it was a waste of time. It limited my job pool initially; this far into my career it really does nothing. I’ve hired some solid bootcamp devs. I’ve seen shitty bootcamp devs. I’ve also seen a bunch of CS masters who have no fucking clue how to ship production code but can wax poetic about algorithm design. Since I don’t run an R&D department, that doesn’t matter 95% of the time.
He's a lunatic because his position is that it won't be a problem. You just train programmers enough that they'll go into the workforce as senior developers.
But that's not how any of the professions he mentioned works:
Software engineering works the same way, you get an internship, then a first job, and both are usually under a senior. In fact, it's not until about 10 years in that I'd consider you an actual senior, and there are levels above that as well.
There's pretty much no industry where you pop out of school at a senior level, there's a reason experience is expected for most roles. I'll only ask about it if you put something interesting, like a relevant project or whatever.
From a hiring perspective, it’s becoming incredibly difficult to weed out AI bullshit. For every one qualified candidate I get, I’ve had to drop five or more in a fucking tech screen
God I'm so afraid to lose job now because I could never survive an interview these days.
I used to shine for things like takehome interview code problems and shit like that, where I had a chance to pause and think a bit and look up definitions and shit.
But those kinds of toy programs are actually the things that AI is actually good at, so now I can only differentiate myself by coding live in front of interviewers and memorizing trivia, both of which I'm terrible at, and don't reflect actual work.
Yeah, it sucks.
What's worse is that we ask a bunch of OOP questions at my company, but I actually hate OOP. I try to work in some FP questions, but those are really hard to ask without using similarly academic language (e.g. describe closure/thunk/partial application and what differentiates them). A lot of people don't know the terminology while knowing the application, because we only cover the terminology in one class in the middle of the curriculum (for OOP; FP was an elective for me), and it's not useful in actual work.
I just want to know if you know what you're doing, and unfortunately, a live coding session usually does the best non at that. Yeah, we're probably missing out on some great devs that just can't perform in an interview, but we're also not having to fire bad devs as much.
Then you should probably have a portfolio of projects.
Between my workday and my family, when am I going to have time to make a portfolio of projects that are complex enough that an AI couldn't have generated them?
Edit: valid criticism of the post is that you have to have a degree to code. That’s bullshit.
Same. I didn't finish even one degree, I'm entirely self taught. I have two prestige positions on my res. Breaking out is incredibly difficult under these circumstances, but once you have one good position that you've held long enough to prove you could do the job, education doesn't matter. You'll probably get at least a phone screening and if you know how to chat with people (not something that comes naturally to everyone), you should likely get a chance to prove yourself in real interviews.
Note: I bombed an interview to an embarrassing degree and got hired by one of the former interviewers when I applied again after leveling up.
As a senior dev who does second round interviews (technical), I certainly ignore education on resumes unless it's an intern or junior position. Experience matters far more to me.
This might be a bit controversial, but all those fields he mentioned do have younger people learning how to do the work. Doctors spend 7 or more years doing doctor work under someone else's watch before they can strike out on their own.
You could call them junior doctors if you like
Nothing controversial there. That person obviously has no idea what they're talking about, as they've clearly never stepped foot on a construction site where junior engineers work alongside senior ones.
The same goes for other professions.
I was first wondering why this even is a LinkedInLunatic, they gave examples that lead me to believe that they were FOR hiring juniors.
in school you learn how to do 2 + 2 and understand why it is 4. its not the actual answer that is the goal, its the understanding of why it is so and how the steps are applied to get the answer. knowledge layered on knowledge happens over years of learning, but eventually you know some stuff about things that people already learnt before you. as the saying goes- this isn't rocket surgery, people!
I'm about to graduate with an M.Sc. in Computer Science - can't wait to be hired as a Senior Engineer!
Lmao.
I don't get what point he is trying to make?
You should do linear regression in excel and call yourself a statistician, is the message, I guess.
insert > scatterplot > click chart > + symbol > trend line > options > linear
ez clap
(this is a joke).
I don't know anyone who is seriously stopping hiring and replacing with AI. Anyone announcing that is just using a hype train to cover poor financials.
It's not that obvious. Corporations are investing heavily in automation in customer relations. There are metrics for how much work had to fall back to humans, because it couldn't be processed by the machine. Managers are motivated to improve on those metrics, and make the humans redundant.
Of course, LLMs are just pure garbage that produce more work for everyone and achieve nothing. Especially in business, they are a great way to reduce efficiency. The users dumb down, believe any bullshit, drop all critical thinking, and the people on the receiving end of their bullshit have to filter even more stupidity than ever.
But you don't understand this as a manager. A piece of code by AI, that produces the same result as a piece of code by a human, or close enough, seem equivalent. Potential side effects are just noise that they don't understand or want to hear about.
Managers also don't understand that AI doesn't scale. If it can write a Python program to calculate prime numbers, it can surely also write something like Netflix, or a payment processor, right?
Then there's exactly what you point out. Other managers claim they're doing it. So there must be something to it.
Once they wasted their budget on renting this technology temporarily, cuts have to be made to ensure the bottom line.
Maybe AI isn't replacing your job, but the stupid investment might cost you the job anyway.
It's also important to realize that you don't require quality work or a quality product to be financially successful as a corporation. The AI industry is the best example itself.
This kind of logic never made sense to me, like: if an AI could build something like Netflix (even if it needed the assistance of a mid software engineer), then it means every indie dev will be able to build a Netflix competitor, bringing the value of Netflix down. Open source tools would quickly reach a level where they’d surpass any closed source software, and would be very user-friendly without much effort.
We’d see LLMs being used to create and improve rapidly infrastructure like compilers, IDEs and build systems that are currently complex and slow, rewrite any slow software into faster languages etc. So many projects that are stalled today for lack of manpower would be flourishing and flooding us with new apps and features in an incredible pace.
I’m yet to see it happen. And that’s because for LLMs to produce anything with enough quality, they need someone who understands what they’re outputting, someone who can add the necessary context in each prompt, who can test it, integrate it into the bigger scheme without causing regressions etc. It’s no simple work and it requires even understanding LLMs’ processing limitations.
LLMs, by definition, can't push the limit. LLM's can only ever produce things that look like what they were trained on. That makes them great for rapidly producing mediocre material, but it also means they will never produce anything complex.
Yeah, people (especially on linkedin) tend to take such BS rather seriously. Facebook said something about replacing engineers with AI, and gumroad said they won't hire anymore (but they don't need a lot of people anyway)
As someone in sourcing who does a lot of "make or buy" decisions, people would be shocked at what people want to replace with AI.
We have a small team of essentially phone jockies that walk users through internal processes and troubleshooting for our janky in house software. They wanted to replace that team with off the shelf AI... No one else uses this software, a lot of information is proprietary, no way AI is going to be able to do that job without specific training.
That's one of a dozen examples where someone tried to ram and AI peg into a square hole.
Any statician worth it's weight is using R, or at least Python (unless it's like a really old statician using spss or SAS). As someone who did interviews for an actuarial intern position, I didn't even asked the candidates if they knew how to use excel, because excel is fucking useless, I asked them about python and pandas.
Idk, my brother used a ton of excel as an actuary. He used other things too, but excel was absolutely part of it, and he made it to VP level in the insurance industry.
There's a still a lot of excel out there being used, but you can't really do a lot of "real" jobs in excel with is 2^20 maximum rows. I don't have a lot of experience myself (I got my degree on 2022), when I interned we used a lot of excel and SAS and I hated it. After that I landed a job where I had the opportunity to write everything from zero in python, and excel is only used to send the results to other teams or clients. In the company I work now, I'm not part of the actuarial team, but in accounting and from the interview it was clear that I was being hired to re write everything from SAS to Python. Sometimes I pass by the actuarial team and I can see them doing chainladder triangles on excel and is kinda sad, because there's a fantastic Python library for that. I'm planning to stay here until everything on the accounting department run on python and then looking for a senior position on the actuarial team to do the same there.
PowerPoint is turing complete in the animations.
You know, if you want to put an interviewee through hell.
PowerPoint is turing complete in the animations.
I promise to only use this knowledge for evil
Thank you.
Hmm. Maybe i should go and volunteer for giving computer classes.
So he's saying that people whose entire qualification are they went through a 2 week boot camp or through a youtube tutorial aren't qualified...? I think? I tend to agree if thats all theyve done, but to be honest a lot of my degree felt like it could have been a 4 hour YT tutorial.
People who get out of uni have no real world experience and should be treated as a juniors though. I've met a lot of people who have book smarts and no idea what to do after theyre in an org. They're weird to work with because you can explain a concept, they'll get it but not be able to apply it or fully see relevance. They're intelligent but lack experience, which seniors provide.
The LinkedIn OP doesn't write clearly, but seems to think junior roles don't do real work. He clearly needs to work in a SOC role to see the difference between a junior and a senior. Lacking experience doesn't mean no meaningful output.
I've hired people who did a bootcamp or whatever, and I tend to ask what else they've done. The good candidates will have an interesting side project and learned a ton outside of the program, and the poor candidates will try to pass classwork off as "personal projects."
We've hired bootcamp people who are way better than people with masters degrees. And I don't just mean faster, I mean they're actually better at the conceptual stuff. Some people just learn better by doing than reading, so they can catch up.
That said, our best "bootcamp engineer" is doing a degree right now while working full time (dude is a beast) to fill in the gaps.
I have met many so many people fresh from university. They studied computer science and they just couldn't code. Like the code quality was abysmal. I literally dealt with people who didn't know how to use git. 6 months later, they were fine. But at the start...
No university has the time to teach you coding.
That guy is acting like universities did.
Yeah I had uni projects with people in the same degree as me (Comp Sci) who straight up said they never learned to code. It baffled me that two people would both leave the degree not knowing the same stuff. I honestly don't know how they were passing classes in some cases.
There's definitely too much knowledge for any one bootcamp, uni course or YT tutorial to teach and experience gaps are hard to identify until they come up. Best thing uni did was teach me how to teach myself, but someone following YT tutorials likely has that skill. That's probably the most important skill to have as someone in tech imo.
There were and are always junior positions in all fields. The other fields are just less self indulgent about the years of experience.
Man doesn't think Residents exist, lmao.
I don't think this guy was ever a junior dumbass. I think he's always been a senior dumbass.
In fields where a college/university degree is a requirement, people start their careers in some kind of junior position anyway.
100% correct. If AI somehow replaces junior devs, someone will have to train them in substitute for paid real-world experience.
It'll be unpaid internships and stuff like that.
I went to university and discovered that unlike that naive view he has of it it is mostly a course in how to tolerate a lot of bullshit from profs and for someone who already taught myself a lot before I got there it was mostly a realization how outdated that whole system has become unless the profs themselves are incredibly motivated (which is relatively rare), the system itself certainly encourages them to do the minimum possible to stay up to date with the material for courses they teach and instead focus on their research.
I have a much longer response, but I’ll try to make a short one. I think there’s a lot more a college degree does (should?) offer/signal, but over the last 50 ish years, that has largely eroded away to just being a professional training program or a gatekeeper to a job. Higher ed in society he mostly turned to social efficiency as its guiding principle instead of several other curricular philosophies. Combine that with the increasing and intense research pressure and it’s the exact situation you describe. Neoliberalism has pushed away long term thinking and risk from corporations, so that burden of risk is taken now by universities (and young people in the form of graduate students) which can be subsidized by government grants. This funding scenario pushes professors to focus on grants and research and to not care about their teaching. It’s not good.
Lmao, love how he thinks linear regression is the epitome of statistics.